@inproceedings{peng-etal-2025-cafe,
    title = "{CAFE}: Retrieval Head-based Coarse-to-Fine Information Seeking to Enhance Multi-Document {QA} Capability",
    author = "Peng, Han  and
      Jiang, Jinhao  and
      Dong, Zican  and
      Zhao, Xin  and
      Fang, Lei",
    editor = "Christodoulopoulos, Christos  and
      Chakraborty, Tanmoy  and
      Rose, Carolyn  and
      Peng, Violet",
    booktitle = "Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing",
    month = nov,
    year = "2025",
    address = "Suzhou, China",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-emnlp/2025.emnlp-main.655/",
    pages = "12977--12989",
    ISBN = "979-8-89176-332-6",
    abstract = "Advancements in Large Language Models (LLMs) have extended their input context length, yet they still struggle with retrieval and reasoning in long-context inputs. Existing methods propose to utilize the prompt strategy and Retrieval-Augmented Generation (RAG) to alleviate this limitation. However, they still face challenges in balancing retrieval precision and recall, impacting their efficacy in answering questions. To address this, we introduce **CAFE**, a two-stage coarse-to-fine method to enhance multi-document question-answering capacities. By gradually eliminating the negative impacts of background and distracting documents, CAFE makes the responses more reliant on the evidence documents. Initially, a coarse-grained filtering method leverages retrieval heads to identify and rank relevant documents. Then, a fine-grained steering method guides attention to the most relevant content. Experiments across benchmarks show that CAFE outperforms baselines, achieving an average SubEM improvement of up to 22.1{\%} and 13.7{\%} over SFT and RAG methods, respectively, across three different models. Our code is available at https://github.com/RUCAIBox/CAFE."
}